AHRQ's Patient Safety Initiative: Building Foundations, Reducing Risk
Appendix 2. The Medical Error Reduction Demonstration Program
Abstracts of the 16 Demonstration Evaluation Grants
Accountability and Health Safety: A Statewide Approach. Georgia Hospital Association Research and Education Foundation—Rollins School of Public Health of Emory University, Atlanta, GA (PI: Kenneth Thorpe, Ph.D.)
The overall aim of this demonstration project is to support improvement in quality of care and patient safety through an effective, replicable, voluntary, peer-review-protected statewide patient safety program in Georgia. The research team is accomplishing this aim by:
- Evaluating the effectiveness of a voluntary error reporting system.
- Evaluating the cost effectiveness of patient safety interventions.
- Developing, implementing, and refining methods to inform patients about safety in general, and injury that occurs as the result of a medical error.
The existing Georgia hospital patient safety program includes a reporting system that was developed to promote a confidential, blame-free environment to facilitate learning from medical errors. In addition to evaluating reporting system effectiveness, researchers are using an iterative process to examine data elements; methods of aggregating, analyzing and reporting data; and approaches to protecting data from unintentional disclosure.
Intensive Care Unit Safety Reporting System (ICUSRS). The Johns Hopkins University, Baltimore, MD (PI: Peter Pronovost, M.D., Ph.D.)
Researchers at the Johns Hopkins University are examining ways to improve patient safety in intensive care units (ICUs) by identifying and eliminating system failures that lead to error in care. They are trying to accomplish this goal through the implementation of a Web-based ICU Safety Reporting System in partnership with the Society for Critical Care Medicine and the Association for Health Care Risk Managers. The researchers are drawing from a variety of experts in error reporting, including staff from the Australian Incident Monitoring Study, the FDA, and human factors experts.
It is estimated that a significant number of the 4 million patients admitted to ICUs in the United States annually suffer a potentially life threatening incident.
Surveillance, Analysis, and Interventions to Improve Patient Safety. Washington University School of Medicine and BJC Healthcare, St. Louis, MO (PI: Victoria Fraser, M.D.)
BJC Healthcare (one of the largest not-for-profit health systems in the United States, with 13 acute care hospitals and over 100 health care delivery sites) and Washington University School of Medicine are studying the best methods to identify:
- Medical errors.
- How to improve reporting of medical errors and adverse events.
- The prevalence of medical errors and their contributing factors, such as medical culture and attitudes.
- The effectiveness of educational and process interventions to reduce medical errors and improve patient safety.
Finally, the researchers are looking at how health care workers communicate with patients and families about medical errors including patients' attitudes and preferences regarding communication about medical errors.
Reporting Systems and Learning: Best Practices. Columbia University, New York, NY (PI: Harold Kaplan, M.D.)
This demonstration project is a consortium effort of two large, geographically and ethnically diverse, integrated health care delivery systems: New York-Presbyterian (the University Hospitals of Columbia and Cornell) and the University of Chicago Hospitals and Health System. The core of this project is a voluntary near-miss medical event reporting system and a State-mandated incident reporting system: the New York Patient Occurrence Reporting and Tracking System. This integrated reporting system is being implemented in New York-Presbyterian Hospital, which will serve as the pilot site for a broader rollout to the constituent hospitals of the New York Presbyterian network and the University of Chicago network.
The specific aims of this proposal are to:
- Expand the medical event reporting system to facilitate reporting of errors by patients and test this approach in an outpatient setting.
- Test novel generalizable, informatics approaches that allow the medical event reporting system to manage and support learning from a large number of reports in an effective and efficient manner.
- Demonstrate the value of reporting by showing its effects on patient safety, organizational culture, and economic outcomes.
- Improve the health care delivery processes and training through the use of failure mode analysis, systems re-design, safety curricula, and simulation based on team training linked to the output of the medical event reporting system.
- Evaluate which types of safety information models of dissemination are valued by consumers, purchasers, policymakers, providers, and regulatory agencies.
- Explore the practical advantages and disadvantages of alternative policy strategies to develop and disseminate such information.
Reporting System to Improve Patient Safety in Surgery. American College of Surgeons, Boston, MA (PI: Shukri Khuri, M.D.)
In 1986, Congress mandated the Department of Veterans' Affairs (VA) to compare its surgical outcomes with the private sector. The National Surgical Quality Improvement Program (NSQIP) was a response to that mandate. Since 1991, the VA NSQIP has been collecting patient risk factors, intraoperative variables, and postoperative outcomes on most patients undergoing major surgery in the VA system. Risk-adjusted outcomes are fed back to the VA surgical programs, so that they can compare their results to other programs in the system. Site visits, medical record reviews, and the dissemination of best surgical practices are offered to high outlier institutions in a program of surgical quality improvement. During this period (1991-2002), there has been a coincident reduction in 30-day operative mortality of 27 percent, in 30-day operative morbidity of 45 percent, and a 45 percent reduction in average postoperative length of stay in the VA system as a whole.
The researchers are trying to determine if similar results can be obtained with the application of the NSQIP reporting system in the private sector. Fourteen non-federal institutions have been recruited and are participating in this study: University of Michigan, University of Kentucky, Emory University, Barnes-Jewish Hospital, Shands Hospital at the University of Florida, Brigham and Women's Hospital, University of California at San Francisco, University of Utah, University of Virginia, Cornell University, Columbia University, University of Maryland, Massachusetts General, and St. Louis University.
Wisconsin Medical Injury Reporting System Demonstration Project. Medical College of Wisconsin, (PI: Peter Layde, M.D., M.Sc.)
The Wisconsin Medical Injury Reporting System demonstration project provides confidential, hospital-specific reports of medical injuries to all non-federal Wisconsin hospitals using State-collected discharge data. A randomized design assigned 127 hospitals to one of three groups including a control group, a feedback only group, and the intervention group which receives both reports and assistance in targeting injury control activities.
Strategies to Identify and Reduce Diabetes Related Medical Errors. Health Partners Research Foundation, MN (PI: Patrick J. O'Connor, M.D., M.P.H.)
Eighty percent of adults with diabetes in the United States are affected by medical errors. The leading cause of error in patients with diabetes is inappropriate or inadequate pharmacologic action resulting in failure to achieve specific clinical aims. These medical errors lead to volumes of preventable adverse events. In the United States, there are $3.9 billion a year of potentially avoidable health care charges. Using a randomized trial, this project is evaluating whether customized direct feedback of medical error information to patients and their physicians reduces medical error rates in diabetics. Eighty primary care physicians and their 1840 diabetic patients affected by medical errors (average of 23 patients with error per physician) will be block randomized to one of four study arms:
- Customized feedback of medical error information to patient only.
- Customized feedback of medical error information to physician only.
- Customized feedback of medical error information to both the patient and their physician.
- Control group with no intervention.
New York State Safety Improvement Demonstration Project. New York State Department of Health, Albany, NY (PI: Wayne Osten, M.P.A.)
The New York State Department of Health is sponsoring three demonstration projects that are studying specific adverse occurrences (i.e., deep vein thrombosis and pulmonary embolism, post-operative heart attacks unrelated to a cardiac procedure, and post-operative wound infections) in the New York Patient Occurrence Report and Tracking System.
Each project will implement new processes of care to reduce the frequency of each of the three adverse occurrences under study, modify physician behavior, and improve the quality and completeness of reporting adverse events in the New York Patient Occurrence Reporting and Tracking System. At baseline, compliance with the existing evidence-based process of care recommendations was 6 to 52 percent.
Applied Strategies for Interventions for Patient Safety. University of Colorado, Aurora, CO (PI: Wilson Pace, M.D.)
Applied Strategies for Improving Patient Safety (ASIPS) is a collaborative effort between the University of Colorado's Department of Family Medicine (UC-DFM) and numerous organizations to analyze the causes and effects of errors in primary care and reduce their incidence. ASIPS is collecting reports of errors or incidents through the Patient Safety Reporting System, a voluntary incident reporting system. The system offers three different ways to confidentially or anonymously report an error: via Web site, telephone hotline, or paper. ASIPS researchers are assessing the specific incidents and trends that may or may not be reported to the UC-DFM, as well as trends occurring in larger populations of practices.
Based on their analyses of primary care errors, the ASIPS team will develop and implement interventions aimed at decreasing recognized errors in primary care. The interventions will be developed by clinicians and staff from practices, with guidance from ASIPS partners, including two insurance organizations, and the State Department of Health. The interventions will be examined using a before-and-after design. The setting for ASIPS is member practices of two practice-based research networks:
- The Colorado Research Network (CaReNet), which focuses on urban minority and underserved primary care populations.
- The High Plains Research Network (HPRN), which focuses on rural, "frontier" primary care practices and hospitals.
The HMO Research Network CERTs Prescribing Safety Program. Harvard Medical School, Harvard Pilgrim Health Care, Cambridge, MA (PI: Richard Platt, M.D.)
Seven AHRQ-funded CERTs, led by the HMO Research Network CERTs (10 HMOs), are collaborating to improve the detection of medication prescribing errors in the ambulatory setting and to develop new uses for this information to improve care. These HMOs, which include staff/group, network, and Independent Practice Associations systems, have nearly 16,000 primary care providers who care for approximately 7 million people in over 1,000 locations.
To detect errors, the researchers are using existing automated pharmacy dispensing information and automated inpatient and outpatient diagnosis and procedure data, plus limited full text record review, to assess the overall frequency of prescribing at variance with either FDA "black box" warnings (a warning of potential serious adverse effects on drug labels) or clinical guidelines. One focus of the research is to detect high risk multiple medication use in the ambulatory setting.
Evaluate the Effects of Massachusetts Reporting Systems. Massachusetts Department of Public Health, Boston, MA (PI: Nancy Ridley, M.S.)
This demonstration project is evaluating the Massachusetts Department of Public Health's (DPH) mandatory hospital reporting system. The study provides an opportunity to evaluate the impact of a unique public-private collaboration established to improve patient safety and reduce medical errors. Given the public nature of data and information contained in the DPH reporting system, researching the issue of providing general, as well as specific information, to patients and families is of critical importance. The Massachusetts DPH is conducting this study in collaboration with the Harvard School of Public Health, The Massachusetts Hospital Association and Coalition for the Prevention of Medical Errors, the Institute for Health Policy at the Massachusetts General Hospital, and the Center for Survey Research at the University of Massachusetts.
Systems Approach for Improving Region-Wide Patient Safety, University of Pittsburgh, PA (PI: Carl Sirio, M.D.)
Preventable adverse events, including nosocomial infections (NIs) and medication errors (MEs), are among the Nation's most pervasive patient safety problems. Establishing effective reporting systems capable of compiling useful information on these events is a necessary but insufficient condition for improving outcomes. Creating the capacity for sustainable change requires region-wide exploration and evaluation of the multiple interrelated systems that transform information into knowledge and learning. The Pittsburgh Regional Healthcare Initiative (PRHI) was created in 1997 to achieve the world's best patient outcomes by identifying and solving systemic problems at the point of patient care, including the elimination of NIs and MEs.
PRHI is a collaboration of leaders from all major health care stakeholder groups in a six-county metro region representing diverse patient populations. Through close collaboration with U.S. Pharmacopeia and the CDC, the 39-hospital system is implementing MEDMARX and components of the CDC's National Nosocomial Infection Surveillance System. All health care facilities participating in this research demonstration have agreed to share their data locally, thus providing an important opportunity for cross-organizational comparisons. Based on shared analysis of regional outcomes data, prevention strategies and interventions are being developed for both NIs and MEs. These will be implemented through a variety of mechanisms including PRHI's existing Center for Shared Learning.
Malpractice Insurers' Medical Error Prevention Study (MIMEPS). Harvard University, Cambridge, MA (PI: David Studdert, Sc.D., L.L.B., M.P.H.)
The study has five goals:
- To harness the potential of malpractice insurance programs to operate as a nationwide error reporting system.
- To identify the most frequent factors contributing to errors found in a focused review of malpractice claims files.
- To test case-control analysis as a method of quantifying the role of various contributing factors to the occurrence of specific errors.
- To assess how the patterns of errors and contributing factors detected through claims file analysis compare to those identified in other reporting systems.
- To use claims data to facilitate design and implementation of a series of targeted patient safety interventions.
To achieve these goals, approximately 2,500 claims files from six large malpractice insurers will be analyzed. Together, these insurers cover approximately 35,000 physicians, 60 acute care hospitals, and more than 100 outpatient facilities in 15 States.
Patient Safety Improvement Using Reporting Systems. Utah Department of Health, Salt Lake City, UT (PI: Scott Williams)
This demonstration project attempts to improve adverse events reporting as well as measure and reduce adverse events in 77 acute care hospitals in Utah and Missouri by examining existing hospital discharge data reporting systems, producing periodic reports on adverse events for hospitals, and conducting regulatory licensure intervention in Missouri and educational intervention in Utah and Missouri. The health departments in Utah and Missouri have statutory authority to collect electronic individual discharge records from all licensed health care facilities and to disseminate analytical results to the facilities and the public in these two States. The States use the administrative data to monitor population morbidity and health care issues such as access, quality, and cost. This project is investigating the expanded use of the hospital discharge data system to improve patient safety.
Addressing Preventable Medication Use Variance in Mississippi. University of Mississippi, MS (PI: Andy Brown)
The University of Mississippi Medical Center (UMMC) is assessing the effectiveness of Mississippi's methods of collecting and using information to reduce medical errors and their impact. Nine sites throughout Mississippi, all of which share the same system for reporting medication errors and, for several academic and administrative functions, report to UMMC's Vice Chancellor for Health Affairs, are participating in this project. The goals of this project are threefold:
- To identify the causes of preventable health care errors and patient injury in health care delivery.
- To develop, demonstrate, and evaluate strategies for reducing errors and improving patient safety throughout Mississippi.
- To disseminate the results of the project's research and the strategies developed throughout the health care industry.
Improved Patient Safety with Information Technology, Indiana University, Regenstrief Medical Center, IN (PI: Mark Overhage)
This project is using a well-established clinical information system (the Regenstrief Medical Record System) to identify "indicators" of errors of omission in ambulatory patient care in 18 practices which are part of a practice-based network. The project is focusing on asthma, chronic obstructive pulmonary disease, congestive heart failure, and lipid management. A clinical focus team, including a specialty physician, general internist, and nurse, will analyze "indicator" data to identify changes in the delivery system that can potentially reduce the "indicators."